Fully Homomorphic Encryption (FHE) enables privacy-preserving Transformer inference, but long-sequence encrypted Transformers quickly exceed single-GPU memory capacity because encoded weights are already large and encrypted activations grow rapidly with sequence length. Multi-GPU execution therefore becomes unavoidable, yet scaling remains challenging because communication is jointly induced by application-level aggregation and encryption-level RNS coupling. Existing approaches either synchronize between devices frequently or replicate encrypted tensors across devices, leading to excessive communication and latency.
We present AEGIS, an Application-Encryption Guided Inference System for scalable long-sequence encrypted Transformer inference on multi-GPU platforms. AEGIS derives device placement from ciphertext dependencies jointly induced by Transformer dataflow and CKKS polynomial coupling, co-locating modulus-coherent and token-coherent data so that communication is introduced only when application dependencies require it, while reordering polynomial operators to overlap the remaining collectives with computation.
On 2048-token inputs, AEGIS reduces inter-GPU communication by up to 57.9% in feed-forward networks and 81.3% in self-attention versus prior state-of-the-art designs. On four GPUs, it achieves up to 96.62% scaling efficiency, 3.86x end-to-end speedup, and 69.1% per-device memory reduction. These results establish coordinated application-encryption parallelism as a practical foundation for scalable homomorphic Transformer inference.
The breakthrough of achieving fully homomorphic encryption sparked enormous studies on where and how to apply homomorphic encryption schemes so that operations can be performed on encrypted data without the secret key while still obtaining the correct outputs. Due to the computational cost, inflated ciphertext size and limited arithmetic operations that are allowed in most encryption schemes, feasible applications of homomorphic encryption are few. While theorists are working on the mathematical and cryptographical foundations of homomorphic encryption in order to overcome the current limitati…
Encrypted control has been introduced to protect controller data by encryption at the stage of computation and communication, by performing the computation directly on encrypted data. In this article, we first review and categorize recent relevant studies on encrypted control. Approaches based on homomorphic encryption, multi-party computation, and secret sharing are introduced, compared, and then discussed with respect to computational complexity, communication load, enabled operations, security, and research directions. We proceed to discuss a current challenge in the application of homomorp…
HE-Guardrail: A Homomorphic Guardrail Against Jailbreak Attacks for Encrypted Large Language Model Inference
Pohang University of Science and Technology (POSTECH)（浦项科技大学）；Inha University（仁荷大学）；Daegu Gyeongbuk Institute of Science and Technology (DGIST)（大邱庆北科学技术院）
We propose a homomorphic search protocol based on quantum homomorphic encryption, in which a client Alice with limited quantum ability can give her encrypted data to a powerful but untrusted quantum server and let the server search for her without decryption. By outsourcing the interactive key-update process to a trusted key center, Alice only needs to prepare and encrypt her original data and to decrypt the ciphered search result in linear time. Besides, we also present a compact and perfectly secure quantum homomorphic evaluation protocol for Cliford circuits, where the decryption key can be…
In this paper, we propose the first homomorphic based proxy re-encryption (HPRE) solution that allows different users to share data they outsourced homomorphically encrypted using their respective public keys with the possibility to process such data remotely. More clearly, this scheme makes possible to switch the public encryption key to another one without the help of a trusted third party. Its originality stands on a method we propose so as to compute the difference between two encrypted data without decrypting them and with no extra communications. Basically, in our HPRE scheme, the two us…
Training Machine Learning Models on Encrypted Data: A Privacy-Preserving Framework using Homomorphic Encryption
ISEP, School of Engineering, Polytechnic of Porto（ISEP工程学院，波尔图理工学院）；GECAD, Research Group on Intelligent Engineering and Computing for Advanced Innovation and Development（GECAD，智能工程与计算高级创新与发展研究组）
Since fully homomorphic encryption and homomorphically encrypted computing preserve algebraic identities such as 2*2=2+2, a natural question is whether this extremely utilitarian feature also sets up cryptographic attacks that use the encrypted arithmetic operators to generate or identify the encryptions of known constants. In particular, software or hardware might use encrypted addition and multiplication to do encrypted division and deliver the encryption of x/x=1. That can then be used to generate 1+1=2, etc, until a complete codebook is obtained. This paper shows that there is no formula o…
In this work, we propose an outsourced Secure Multilayer Perceptron (SMLP) scheme where privacy and confidentiality of both the data and the model are ensured during the training and the classification phases. More clearly, this SMLP : i) can be trained by a cloud server based on data previously outsourced by a user in an homomorphically encrypted form; ii) its parameters are homomorphically encrypted giving thus no clues to the cloud; and iii) it can also be used for classifying new encrypted data sent by the user returning him the encrypted classification result encrypted. The originality of…
Recent advances in cryptography promise to enable secure statistical computation on encrypted data, whereby a limited set of operations can be carried out without the need to first decrypt. We review these homomorphic encryption schemes in a manner accessible to statisticians and machine learners, focusing on pertinent limitations inherent in the current state of the art. These limitations restrict the kind of statistics and machine learning algorithms which can be implemented and we review those which have been successfully applied in the literature. Finally, we document a high performance R …
The database community, at least for the last decade, has been grappling with querying encrypted data, which would enable secure database as a service solutions. A recent breakthrough in the cryptographic community (in 2009) related to fully homomorphic encryption (FHE) showed that arbitrary computation on encrypted data is possible. Successful adoption of FHE for query processing is, however, still a distant dream, and numerous challenges have to be addressed. One challenge is how to perform algebraic query processing of encrypted data, where we produce encrypted intermediate results and oper…
New cryptographic techniques such as homomorphic encryption (HE) allow computations to be outsourced to and evaluated blindfolded in a resourceful cloud. These computations often require private data owned by multiple participants, engaging in joint evaluation of some functions. For example, Genome-Wide Association Study (GWAS) is becoming feasible because of recent proliferation of genome sequencing technology. Due to the sensitivity of genomic data, these data should be encrypted using different keys. However, supporting computation on ciphertexts encrypted under multiple keys is a non-trivi…
Falcon: Accelerating Homomorphically Encrypted Convolutions for Efficient Private Mobile Network Inference
School of Integrated Circuits, Peking University（北京大学集成电路学院）；Institute for Artificial Intelligence, Peking University（北京大学人工智能研究院）；Institute of Electronic Design Automation, Peking University（北京大学电子设计自动化研究所）；Beijing Advanced Innovation Center for Integrated Circuits（北京集成电路先进制造创新中心）
Deep Learning as a Service (DLaaS) stands as a promising solution for cloud-based inference applications. In this setting, the cloud has a pre-learned model whereas the user has samples on which she wants to run the model. The biggest concern with DLaaS is user privacy if the input samples are sensitive data. We provide here an efficient privacy-preserving system by employing high-end technologies such as Fully Homomorphic Encryption (FHE), Convolutional Neural Networks (CNNs) and Graphics Processing Units (GPUs). FHE, with its widely-known feature of computing on encrypted data, empowers a wi…
École Supérieure en Informatique, Sidi Bel Abbès（西迪贝勒阿巴斯高等信息学院）；OpenMined
Machine learning algorithms have achieved remarkable results and are widely applied in a variety of domains. These algorithms often rely on sensitive and private data such as medical and financial records. Therefore, it is vital to draw further attention regarding privacy threats and corresponding defensive techniques applied to machine learning models. In this paper, we present TenSEAL, an open-source library for Privacy-Preserving Machine Learning using Homomorphic Encryption that can be easily integrated within popular machine learning frameworks. We benchmark our implementation using MNIST…
We consider an architecture of confidential cloud-based control synthesis based on Homomorphic Encryption (HE). Our study is motivated by the recent surge of data-driven control such as deep reinforcement learning, whose heavy computational requirements often necessitate an outsourcing to the third party server. To achieve more flexibility than Partially Homomorphic Encryption (PHE) and less computational overhead than Fully Homomorphic Encryption (FHE), we consider a Reinforcement Learning (RL) architecture over Leveled Homomorphic Encryption (LHE). We first show that the impact of the encryp…
Fully homomorphic encryption (FHE) allows computations over encrypted data. This technique makes privacy-preserving cloud computing a reality. Users can send their encrypted sensitive data to a cloud server, get encrypted results returned and decrypt them, without worrying about data breaches. This project report presents a homomorphic instruction set emulator, CryptoEmu, that enables fully homomorphic computation over encrypted data. The software-based instruction set emulator is built upon an open-source, state-of-the-art homomorphic encryption library that supports gate-level homomorphic ev…
Rochester Institute of Technology（罗切斯特理工学院）；Taif University（塔伊夫大学）
Homomorphic encryption (HE) is a promising cryptographic technique for enabling secure collaborative machine learning in the cloud. However, support for homomorphic computation on ciphertexts under multiple keys is inefficient. Current solutions often require key setup before any computation or incur large ciphertext size (at best, grow linearly to the number of involved keys). In this paper, we proposed a new approach that leverages threshold and multi-key HE to support computations on ciphertexts under different keys. Our new approach removes the need for key setup between each client and th…
Following a sequence of hardware designs for a fully homomorphic crypto-processor - a general purpose processor that natively runs encrypted machine code on encrypted data in registers and memory, resulting in encrypted machine states - proposed by the authors in 2014, we discuss a working prototype of the first of those, a so-called `pseudo-homomorphic' design. This processor is in principle safe against physical or software-based attacks by the owner/operator of the processor on user processes running in it. The processor is intended as a more secure option for those emerging computing parad…
Catalano and Fiore propose a scheme to transform a linearly-homomorphic encryption into a homomorphic encryption scheme capable of evaluating quadratic computations on ciphertexts. Their scheme is based on the linearly-homomorphic encryption (such as Goldwasser-Micali, Paillier and ElGamal) and need to perform large integer operation on servers. Then, their scheme have numerous computations on the servers. At the same time, their scheme cannot verify the computations and cannot evaluate more than degree-4 computations. To solve these problems, we no longer use linearly-homomorphic encryption w…
Cloud computing emerges as an attractive solution that can be delegated to store and process confidential data. However, several security risks are encountered with such a system as the securely encrypted data should be decrypted before processing them. Therefore, the decrypted data is susceptible to reading and alterations. As a result, processing encrypted data has been a research subject since the publication of the RSA encryption scheme in 1978. In this paper we present a relational database system based on homomorphic encryption schemes to preserve the integrity and confidentiality of the…
Fully homomorphic encryption (FHE) allows anyone to perform computations on encrypted data, despite not having the secret decryption key. Since the Gentry's work in 2009, the primitive has interested many researchers. In this paper, we stress that any computations performed on encrypted data are constrained to the encrypted domain (finite fields or rings). This restriction makes the primitive useless for most computations involving common arithmetic expressions and relational expressions. It is only applicable to the computations related to modular arithmetic. We want to reaffirm that cryptogr…
Fully homomorphic encryption (FHE) enables a simple, attractive framework for secure search. Compared to other secure search systems, no costly setup procedure is necessary; it is sufficient for the client merely to upload the encrypted database to the server. Confidentiality is provided because the server works only on the encrypted query and records. While the search functionality is enabled by the full homomorphism of the encryption scheme. For this reason, researchers have been paying increasing attention to this problem. Since Akavia et al. (CCS 2018) presented a framework for secure sear…
With the rapid development of cloud computing, the privacy security incidents occur frequently, especially data security issues. Cloud users would like to upload their sensitive information to cloud service providers in encrypted form rather than the raw data, and to prevent the misuse of data. The main challenge is to securely process or analyze these encrypted data without disclosing any useful information, and to achieve the rights management efficiently. In this paper, we propose the encrypted data processing protocols for cloud computing by utilizing additively homomorphic encryption and …
In this paper we compare the performance of various homomorphic encryption methods on a private search scheme that can achieve $k$-anonymity privacy. To make our benchmarking fair, we use open sourced cryptographic libraries which are written by experts and well scrutinized. We find that Goldwasser-Micali encryption achieves good enough performance for practical use, whereas fully homomorphic encryptions are much slower than partial ones like Goldwasser-Micali and Paillier.
This paper proposes a fully homomorphic encryption encapsulated difference expansion (FHEE-DE) scheme for reversible data hiding in encrypted domain (RDH-ED). In the proposed scheme, we use key-switching and bootstrapping techniques to control the ciphertext extension and decryption failure. To realize the data extraction directly from the encrypted domain without the private key, a key-switching based least-significant-bit (KS-LSB) data hiding method has been designed. In application, the user first encrypts the plaintext and uploads ciphertext to the server. Then the server performs data hid…
The need for privacy-preserving analytics is higher than ever due to the severity of privacy risks and to comply with new privacy regulations leading to an amplified interest in privacy-preserving techniques that try to balance between privacy and utility. In this work, we present an efficient method for Text Classification while preserving the privacy of the content using Fully Homomorphic Encryption (FHE). Our system (named \textbf{Priv}ate \textbf{F}ast \textbf{T}ext (PrivFT)) performs two tasks: 1) making inference of encrypted user inputs using a plaintext model and 2) training an effecti…
Computationally demanding tasks are typically calculated in dedicated data centers, and real-time visualizations also follow this trend. Some rendering tasks, however, require the highest level of confidentiality so that no other party, besides the owner, can read or see the sensitive data. Here we present a direct volume rendering approach that performs volume rendering directly on encrypted volume data by using the homomorphic Paillier encryption algorithm. This approach ensures that the volume data and rendered image are uninterpretable to the rendering server. Our volume rendering pipeline…
Convolutional neural networks (CNNs) have enabled significant performance leaps in medical image classification tasks. However, translating neural network models for clinical applications remains challenging due to data privacy issues. Fully Homomorphic Encryption (FHE) has the potential to address this challenge as it enables the use of CNNs on encrypted images. However, current HE technology poses immense computational and memory overheads, particularly for high-resolution images such as those seen in the clinical context. We present CaRENets: Compact and Resource-Efficient CNNs for high per…
School of Computer Science and Statistics（计算机科学与统计学院）
Credit card fraud is a problem continuously faced by financial institutions and their customers, which is mitigated by fraud detection systems. However, these systems require the use of sensitive customer transaction data, which introduces both a lack of privacy for the customer and a data breach vulnerability to the card provider. This paper proposes a system for private fraud detection on encrypted transactions using homomorphic encryption. Two models, XGBoost and a feedforward classifier neural network, are trained as fraud detectors on plaintext data. They are then converted to models whic…
Barcelona Supercomputing Center (BSC)（巴塞罗那超级计算中心）；Intel Corporation（英特尔公司）
The proliferation of machine learning services in the last few years has raised data privacy concerns. Homomorphic encryption (HE) enables inference using encrypted data but it incurs 100x-10,000x memory and runtime overheads. Secure deep neural network (DNN) inference using HE is currently limited by computing and memory resources, with frameworks requiring hundreds of gigabytes of DRAM to evaluate small models. To overcome these limitations, in this paper we explore the feasibility of leveraging hybrid memory systems comprised of DRAM and persistent memory. In particular, we explore the rece…
Lehigh University（理海大学）；Microsoft（微软）；Florida International University（佛罗里达国际大学）
Recently cloud-based graph convolutional network (GCN) has demonstrated great success and potential in many privacy-sensitive applications such as personal healthcare and financial systems. Despite its high inference accuracy and performance on cloud, maintaining data privacy in GCN inference, which is of paramount importance to these practical applications, remains largely unexplored. In this paper, we take an initial attempt towards this and develop $\textit{CryptoGCN}$--a homomorphic encryption (HE) based GCN inference framework. A key to the success of our approach is to reduce the tremend…
Homomorphic encryption has been an area of study in classical computing for decades. The fundamental goal of homomorphic encryption is to enable (untrusted) Oscar to perform a computation for Alice without Oscar knowing the input to the computation or the output from the computation. Alice encrypts the input before sending it to Oscar, and Oscar performs the computation directly on the encrypted data, producing an encrypted result. Oscar then sends the encrypted result of the computation back to Alice, who can decrypt it. We describe an approach to homomorphic encryption for quantum annealing …
In data-driven predictive cloud control tasks, the privacy of data stored and used in cloud services could be leaked to malicious attackers or curious eavesdroppers. Homomorphic encryption technique could be used to protect data privacy while allowing computation. However, extra errors are introduced by the homomorphic encryption extension to ensure the privacy-preserving properties, and the real number truncation also brings uncertainty. Also, process and measure noise existed in system input and output may bring disturbance. In this work, a data-driven predictive cloud controller is develope…
Fully homomorphic encryption enables arbitrary computation on encrypted data without decrypting the data. Here it is studied in the context of quantum information processing. Based on universal quantum circuit, we present a quantum fully homomorphic encryption (QFHE) scheme, which permits arbitrary quantum transformation on an encrypted data. The QFHE scheme is proved to be perfectly secure. In the scheme, the decryption key is different from the encryption key, however, the encryption key cannot be public. Moreover, the evaluate algorithm of the scheme is independent of the encryption key, so…
